Research On Distributed Allocation System Based On Multi-source Model Transformation Task Optimization Scheduling

With the rapid development of information technology and the popularity of Internetrelated technologies,cloud services as an emerging service model are spreading and developing at an exponential rate.However,the explosive growth of related data and number of cloud tasks not only bring great pressure to the system background,but also affect the overall performance and user experience of the whole system.Therefore,the architecture of the cloud service system background gradually evolves to distributed,reducing system pressure through distributed design and deployment,and improving the interactive experience of system.Therefore,achieving distributed processing of batch tasks and combining characteristics of task and processor to complete reasonable task scheduling in a distributed environment,it becomes the key to solve the problem of cloud service systems load expansion.In this paper,based on the problems encountered in the actual project,aiming at the task characteristics and requirements of the multi-source 3D model online conversion display platform,research and design a distributed allocation system dedicated to the transformation of the model,to solve the problem of excessive cloud task accumulation and slow user response.According to the distributed basic principle to built a distributed task processing system,and construct the problem model of multi-source model task scheduling and turn it into a goal optimization problem.The task scheduling algorithm for the problem model is designed to solve the optimal target.The specific work is as follows:Firstly,analyzing the specific business process and characteristics of the distributed allocation system,and using the optimal time-span of completing the single batch of tasks as the scheduling target,and describing the task scheduling problem in the distributed allocation system formally.Then based on the traditional genetic algorithm,this paper designs and improves a new hybrid algorithm GBCA algorithm.,to solve the above distributed task scheduling problem and obtain the optimal or suboptimal task scheduling plan in a limited time.Secondly,according to the actual application scenario to built distributed processing system.Firstly,analyzing the design goals and functional requirements of the system and designing overall architecture of the system.Then according to the business logic,the distributed allocation system is divided into three modules,including the task startup module,the scheduling execution module,and the function interaction module,and each module is designed and studied in detail.Finally,the experimental platform was built to complete the deployment of the distributed allocation system in processors,and the function and performance of the system were tested.Results of the system test show that the distributed allocation system can complete the expected function successfully.And the task scheduling model and GBCA algorithm can reduce the timespan of single batch task effectively,help the system overcome the problem of excessive task accumulation and poor interactive experience.